AI for Connectivity Hackathon II: Building Resilient Networks 

Start

HelixAI is committed to develop a highly effective application for network health optimization.

HelixAI

Muhammad Hashir Ehtisham

MominaKanwal

Aditi Prabhakar

This project "NetPulse AI" focuses on network health monitoring and fault prediction using a Raspberry Pi and AI-powered analysis. The Raspberry Pi collects real-time system performance data—including CPU usage, temperature, signal strength, and packet loss—via sensors and serial communication. A VBScript automates data logging, storing readings in an Excel file without user intervention. The script detects anomalies and logs detailed error messages to help diagnose potential failures. This setup enables continuous monitoring without manual oversight, ensuring that network performance metrics are recorded efficiently.

Once the data is logged, users can upload the Excel file to an AI-powered chatbot hosted on Hugging Face. The chatbot, built using Gradio, Pandas, and Zephyr-7B-Beta, processes the data to provide insights under three key sections: Future Performance Prediction, Risk Analysis and Potential Issues, and Preventive Actions and Recommendations. The AI detects patterns, predicts potential failures, and offers optimization strategies for network stability. Users can also download a detailed report and interact with the chatbot for additional troubleshooting and performance improvement suggestions.

NetPulse AI is an innovative project designed  predict network failures and optimize performance using AI-driven analysis of real-time data from Raspberry Pi.

"HelixAI team on AI for Connectivity Hackathon II: Building Resilient Networks Hackathon"

Team Idea

Submission